Vision Technology Research Articles

Development of Computer Vision-based Visual Feedback Assistive Technology for Fine Motor Improvement in Children with Cerebral Palsy

Cerebral Palsy (CP) is a neuromotor condition that affects fine motor development, causing barriers to daily activities such as writing and grasping. This research develops an assistive technology that utilizes motion sensors to provide real-time visual feedback, which helps children understand and correct movements more effectively. The technology improves movement precision and coordination in motor skills, learner engagement, and structured monitoring of skill improvement. The uniqueness of this research lies in integrating computer vision technology with an interactive design that allows CP children to perform motor exercises and engage them in a more personalized and adaptive learning process. The Research and Development (R&D) method was used in the development, involving validating material experts, media experts, and educational practitioners. The results of media development show that assistive technology is feasible to use in improving the fine motor skills of children with cerebral palsy. Material experts stated an average value of 3.92, which was accumulated in percentage to 98.21%; media experts obtained results of 3.71, which was collected in percentage to 92.85%; and expert practitioners stated an average value of 3.93, which was accumulated in percentage to 98.43%. The assistive technology developed is worth testing in improving the fine motor skills of cerebral palsy children. This technology is expected to solve motor and coordination training issues at school and home. With a more interactive and fun approach, CP children can be more motivated to practice so that their motor skills can develop better.

A roadmap for governing AI: technology governance and power-sharing liberalism

Abstract This paper aims to provide a roadmap for governing AI. In contrast to the reigning paradigms, we argue that AI governance should be not merely a reactive, punitive, status-quo-defending enterprise, but rather the expression of an expansive, proactive vision for technology—to advance human flourishing. Advancing human flourishing in turn requires democratic/political stability and economic empowerment. To accomplish this, we build on a new normative framework that will give humanity its best chance to reap the full benefits, while avoiding the dangers, of AI. This new framework of “Power-Sharing Liberalism” is a philosophy that restores protections of positive liberties to liberalism. As we deploy it here, it helps shape a more comprehensive (and we would contend, more accurate) understanding of both risk and opportunity introduced by AI. To lay out how Power-Sharing Liberalism can be applied to AI governance, we take four steps. First, we define central concepts in the field of AI governance, disambiguating between forms of technological harms and risks. Second, we review current normative frameworks around the globe and argue that Power-Sharing Liberalism is a better fit for governing AI. Third, we walk through the six governance tasks that should be accomplished by any governance framework and analyze them through the Power-Sharing Liberalism framework. Based on that analysis, we make 17 recommendations for the governance of AI—including transformative investments in public goods, personnel, and the sustainability of democracy itself. Finally, we discuss concrete proposals for implementing those recommendations.

DPD-YOLO: dense pineapple fruit target detection algorithm in complex environments based on YOLOv8 combined with attention mechanism

With the development of deep learning technology and the widespread application of drones in the agricultural sector, the use of computer vision technology for target detection of pineapples has gradually been recognized as one of the key methods for estimating pineapple yield. When images of pineapple fields are captured by drones, the fruits are often obscured by the pineapple leaf crowns due to their appearance and planting characteristics. Additionally, the background in pineapple fields is relatively complex, and current mainstream target detection algorithms are known to perform poorly in detecting small targets under occlusion conditions in such complex backgrounds. To address these issues, an improved YOLOv8 target detection algorithm, named DPD-YOLO (Dense-Pineapple-Detection YOU Only Look Once), has been proposed for the detection of pineapples in complex environments. The DPD-YOLO model is based on YOLOv8 and introduces the attention mechanism (Coordinate Attention) to enhance the network’s ability to extract features of pineapples in complex backgrounds. Furthermore, the small target detection layer has been fused with BiFPN (Bi-directional Feature Pyramid Network) to strengthen the integration of multi-scale features and enrich the extraction of semantic features. At the same time, the original YOLOv8 detection head has been replaced by the RT-DETR detection head, which incorporates Cross-Attention and Self-Attention mechanisms that improve the model’s detection accuracy. Additionally, Focaler-IoU has been employed to improve CIoU, allowing the network to focus more on small targets. Finally, high-resolution images of the pineapple fields were captured using drones to create a dataset, and extensive experiments were conducted. The results indicate that, compared to existing mainstream target detection models, the proposed DPD-YOLO demonstrated superior detection performance for pineapples in situations where the background is complex and the targets are occluded. The mAP@0.5 reached 62.0%, representing an improvement of 6.6% over the original YOLOv8 algorithm, Precision increased by 2.7%, Recall improved by 13%, and F1-score rose by 10.3%.

Ultralow Turn‐On Voltage Organic Upconversion Devices for High‐Resolution Imaging Based on Near‐Infrared Homotandem Photodetector

AbstractIn recent years, infrared visualization technology has attracted significant attention, and organic upconversion devices (OUDs) have become a reliable infrared imaging strategy due to the superiority of low cost, simple structure, and large‐area imaging. Nonetheless, the power consumption of OUDs is mostly high, so effectively reducing the turn‐on voltage (Von) has emerged as a crucial aspect in optimizing OUD performance, alongside the urgent need to enhance imaging resolution. This work presents an efficient OUD integrating a homotandem photodetector and an organic light‐emitting diode to reduce the Von of the OUD since the homotandem detection unit can generate nearly twice the open‐circuit voltage (Voc) compared to the single detection unit. The device reaches a noteworthy Von of 0.64 V, which is the lowest Von reported for OUDs, accompanied by a high upconversion efficiency of 10.64% and a wide luminance linear dynamic range (L‐LDR) of 82.94 dB. In addition, high‐quality bioimaging with wide operating voltages is demonstrated, achieving imaging resolution up to 5799 pixels per inch (ppi). This work demonstrates an effective technical approach to reduce the power consumption of OUDs and facilitate future bioimaging applications.

Attendance System for College Hostels Using Facial Recognition

Face detection is a computer vision technology designed to identify and locate human faces in digital images. It is a specialized application of object detection, which involves identifying instances of specific semantic objects, such as humans, buildings, or vehicles, in images and videos. With advancements in technology, face detection has become increasingly significant in fields like photography, security, and marketing. This report presents an efficient approach to detecting and recognizing human faces using OpenCV and Python. It explores the pivotal role of machine learning in computer science and its application in facial detection through various OpenCV libraries. Furthermore, the report proposes a system for real-time human face detection, leveraging the integration of machine learning techniques with OpenCV and Python.

Emerging Trends on Hybrid Artificial Intelligence Framework for Transforming Heavy Machinery with Robots and Energy Efficiency

Today, heavy machinery can be equipped with robots operated by artificial intelligence (AI) to streamline operations, decrease human labor, and increase efficiency. Intelligent technologies like these can boost productivity by carrying out routine tasks with extreme accuracy. Since robots reduce risks to human workers, robots with AI algorithms and sensors could be used in hazardous environments. In certain contexts, including building site management, mining, or handling hazardous materials, safety must take precedence above everything else. Analytics from this study, aided by AI, have opened the door to predictive maintenance procedures. To detect equipment problems, the Hybrid Artificial Intelligence Framework (HAIF) examines sensor data in conjunction with historical patterns. The objective is to avoid expensive failures and downtime. Optimization of equipment utilization, energy consumption, and fuel consumption can be achieved by applying AI-driven machine learning algorithms. Several advantages arise from energy waste, including cost avoidance and enhanced operational efficacy. Robots can do exact alignments, measurements, and operations with the help of AI vision technologies. This precision significantly impacts the manufacturing, agricultural, and construction industries. Artificial intelligence has the potential to optimize the allocation of resources, the use of heavy equipment, and supply networks by evaluating massive amounts of data.

Research on the Application of Computer Vision Technology in Sports Mechanics Analysis

With the continuous development of computer vision technology and related hardware devices, motion mechanics analysis is gradually being applied in sports training. However, due to the complex characteristics of motion mechanics and irregular movements, motion mechanics analysis still has limitations in practical application scenarios. This article constructs a multi convolutional 3D CNN model that combines BN algorithm, dropout technique, and spatial pyramid pooling technique. Different features are used as inputs for 3D CNN models tested on video datasets. The experimental results show that combining the “BM+OFM+three FDF” features as model inputs can achieve high recognition accuracy. Therefore, the 3D CNN model constructed in this article can effectively improve the accuracy of motion mechanics recognition and has good application value.

Comprehensive evaluation of the ramification patterns of hepatic vascular anatomy based on three-dimensional visualization technology.

The liver segmentation method proposed by Couinaud is widely accepted by surgeons because of its convenience and practicality. However, this conventional eight-segment classification does not reflect realistic details of the liver and thus requires further adjustments to promote improvements in surgical strategies. This study aimed to explore the ramification patterns of the hepatic vasculature comprehensively. A total of 197 eligible patients meeting the study criteria were enrolled for three-dimensional reconstruction analysis. In the left hemiliver, the portal vein bifurcated into P2 and umbilical portion (UP) in 172 (98.3%) patients. The P4b of 103 patients (103/172, 59.9%) whose P4b branched from the right horn of the left portal vein (LPV) diverged from the main trunk of the UP. In the right paramedian sector (RPMS), the entire portal trunk directly bifurcates into P8vent and P8dor. Simple branching of P5 off the trunk of the RPMS was observed in 78 patients (78/130, 60%). The anterior fissure vein (AFV) was identified in 86 (86/148, 58.1%) patients. V8d entered the right hepatic vein (RHV) in all the patients. In 75.3% (113/150) of all the patients, V5d joined the RHV. In the right lateral sector (RLS), more than half (71/133, 53.4%) of the patients had an arch-like type. We summarize different patterns of liver vascular branches, providing a valuable reference for clinical surgery and liver transplantation. Cranio-caudal segmentation is more common than ventral-dorsal segmentation. The AFV can be regarded as a reliable anatomical landmark for subsegmentation in segment 8. In addition, the absence of AFV was associated with the P8 pattern.

Real-time detection and monitoring of public littering behavior using deep learning for a sustainable environment

With the global population surpassing 8 billion, waste production has skyrocketed, leading to increased pollution that adversely affects both terrestrial and marine ecosystems. Public littering, a significant contributor to this pollution, poses severe threats to marine life due to plastic debris, which can inflict substantial ecological harm. Additionally, this pollution jeopardizes human health through contaminated food and water sources. Given the annual global plastic consumption of approximately 475 million tons and the pervasive issue of public littering, addressing this challenge has become critically urgent. The Surveillance and Waste Notification (SAWN) system presents an innovative solution to combat public littering. Leveraging surveillance cameras and advanced computer vision technology, SAWN aims to identify and reduce instances of littering. Our study explores the use of the MoViNet video classification model to detect littering activities by vehicles and pedestrians, alongside the YOLOv8 object detection model to identify individuals responsible through facial recognition and license plate detection. Collecting appropriate data for littering detection presented significant challenges due to its unavailability. Consequently, project members simulated real-life littering scenarios to gather the required data. This dataset was then used to train different models, including LRCN, CNN-RNN, and MoViNets. After extensive testing, MoViNets demonstrated the most promising results. Through a series of experiments, we progressively improved the model’s performance, achieving accuracy rates of 93.42% in the first experiment, 95.53% in the second, and ultimately reaching 99.5% in the third experiment. To detect violators’ identities, we employed YOLOv8, trained on the KSA vehicle plate dataset, achieving 99.5% accuracy. For face detection, we utilized the Haar Cascade from the OpenCV library, known for its real-time performance. Our findings will be used to further enhance littering behavior detection in future developments.

AI- and Deep Learning-Powered Driver Drowsiness Detection Method Using Facial Analysis

The significant number of road traffic accidents caused by fatigued drivers presents substantial risks to the public’s overall safety. In recent years, there has been a notable convergence of intelligent cameras and artificial intelligence (AI), leading to significant advancements in identifying driver drowsiness. Advances in computer vision technology allow for the identification of driver drowsiness by monitoring facial expressions such as yawning, eye movements, and head movements. These physical indications, together with assessments of the driver’s physiological condition and behavior, aid in assessing fatigue and lowering the likelihood of drowsy driving-related incidents. This study presents an extensive variety of meticulously designed algorithms that were thoroughly analyzed to assess their effectiveness in detecting drowsiness. At the core of this attempt lay the essential concept of feature extraction, an efficient technique for isolating facial and ocular regions from a particular set of input images. Following this, various deep learning models, such as a traditional CNN, VGG16, and MobileNet, facilitated detecting drowsiness. Among these approaches, the MobileNet model was a valuable choice for drowsiness detection in drivers due to its real-time processing capability and suitability for deployment in resource-constrained environments, with the highest achieved accuracy of 92.75%.

A Reconstruction of the Shrine of the Prophet Nahum: An Analysis of 3D Documentation Methods and Data Transfer Technology for Virtual and Augmented Realities

This article focuses on modern methods of documentation and visualization for a historic object. Digital photogrammetry and terrestrial laser scanning (TLS), which are essential tools for documenting cultural heritage in view of their rapid development in recent years, were used, compared, and analyzed. Furthermore, the use of available 3D computer graphics technologies for visualization is described and an optimal procedure for converting the object into VR and AR is proposed and implemented. The technologies presented in this article were tested within the context of a project on the reconstruction of the shrine of the Prophet Nahum in the city of Alqosh in northern Iraq, taking the shrine as a case study. Funded by ARCH Int. and provided by GemaArt Int., the restoration project started in 2018 and was completed in 2021. The ongoing documentation was prepared by the CTU and it used the materials for research purposes. Accurate documentation using photogrammetry, drones, and TLS was key to the restoration. Leica BLK360, Faro Focus S150, and GeoSlam laser scanners were used, as well as photogrammetric methods. In particular, the documentation process involved the creation of 3D textured models from the photogrammetry, which were compared to the TLS data to ensure accuracy. These models were necessary to track changes during the reconstruction phases and to calculate the volumes of rubble removed and materials added. Our data analysis revealed significant differences between the construction logs and the analysis of the accurate 3D models; the results showed an underestimation of the displaced material statements by 13.4% for removed material and 4.6% for added material. The use of heat maps and volumetric analyses helped to identify areas of significant change that guided the reconstruction and documented significant changes to the building for the investor. These findings are important for use in the construction industry with respect to historic sites as well as for further research focused on visualization using VR (virtual reality) and AR (augmented reality). The conversion of existing 3D models into VR and AR is rapidly evolving and significant progress was made during this project. The Unreal Engine (UE) game engine was used. Despite the significantly improved performance of the new UE 5 version, the data for conversion to VR and AR needs to be decimated to reduce the amount—in our case, this was by up to 90%. The quality appearance of the objects is then ensured by textures. An important outcome of this part of the research was the debugged workflow developed to optimize the 3D models for VR, which was essential for creating a virtual museum that shows the restoration process.

Applications and challenges of artificial intelligence-driven 3D vision in biomedical engineering: A biomechanics perspective

This paper explores the applications and challenges of artificial intelligence (AI)-driven 3D vision technology in biomedical engineering, with a specific focus on its integration with biomechanics. 3D vision technology offers richer spatial information compared to traditional 2D imaging and is increasingly applied in fields like medical image analysis, surgical navigation, lesion detection, and biomechanics. In biomechanics, AI-driven 3D vision is used for analyzing human movement, modeling musculoskeletal systems, and assessing joint biomechanics. However, challenges persist, including image quality, computational resource demands, data privacy, and algorithmic bias. This paper reviews the development of 3D vision technology and AI, discusses its applications in biomedicine and biomechanics, and addresses the key technical obstacles, offering insights into the future development of these technologies in the context of biomedical and biomechanical research.

On methods of setting the coordinates of the mouse pointer in the case of alternative options for constructing human-computer interaction

Objective. The aim of the study is to compare and generalize different methods of contactless setting of computer mouse pointer coordinates. Method. To create a contactless human-machine interface, computer vision technologies are used to select a control point on a video camera frame, with the help of which the computer cursor is moved on the user's monitor. General logical methods of scientific research, namely, generalization, are used. An analysis of the considered cursor coordinate control systems is carried out with the allocation of common parts for further generalization and description of formulas for setting the coordinates of the computer mouse cursor. Scientific works with absolute positioning of the computer cursor and works with control of the cursor position shift relative to its coordinates on the previous frame were taken for analysis. Result. As a result of the analysis, strengths and weaknesses of each of the approaches under consideration were indicated. Absolute positioning requires less time to set the cursor to the desired position. It is determined that cursor shift control is not so demanding on the accuracy of calculating the coordinates of the control point. Formulas are given that generalize the methods of moving the computer cursor. Conclusion. The proposed approach to calculating new coordinates of the mouse pointer allows implementing both presented approaches to positioning the computer cursor in the software system, since each of them has advantages depending on the specific situation.

Application of 3D Virtual Digital Visualization Technology in the Simulation and Modeling of Cross-Sea Network Engineering

This paper analyzes the importance of cross-sea interconnection projects and delves into the application of 3D virtual digital visualization technology in the simulation of these projects. The system employs OpenGL technology to render 3D images and supports environmental configuration through dynamic link libraries. The internal design of the system is clear, featuring functionalities such as multi-scene dynamic focusing, multi-data port linkage, BIM information display, event tracing, and process simulation. Compared with traditional construction techniques, 3D virtual visualization technology exhibits significant advantages in graphic rendering, scheme selection, collaboration, and construction methods. The research presented in this paper provides valuable references for optimizing cross-sea interconnection projects and other major infrastructure constructions, demonstrating the immense potential of 3D virtual visualization technology in enhancing construction efficiency and reliability.

Deep learning and computer vision in plant disease detection: a comprehensive review of techniques, models, and trends in precision agriculture

Plant diseases cause significant damage to agriculture, leading to substantial yield losses and posing a major threat to food security. Detection, identification, quantification, and diagnosis of plant diseases are crucial parts of precision agriculture and crop protection. Modernizing agriculture and improving production efficiency are significantly affected by using computer vision technology for crop disease diagnosis. This technology is notable for its non-destructive nature, speed, real-time responsiveness, and precision. Deep learning (DL), a recent breakthrough in computer vision, has become a focal point in agricultural plant protection that can minimize the biases of manually selecting disease spot features. This study reviews the techniques and tools used for automatic disease identification, state-of-the-art DL models, and recent trends in DL-based image analysis. The techniques, performance, benefits, drawbacks, underlying frameworks, and reference datasets of more than 278 research articles were analyzed and subsequently highlighted in accordance with the architecture of computer vision and deep learning models. Key findings include the effectiveness of imaging techniques and sensors like RGB, multispectral, and hyperspectral cameras for early disease detection. Researchers also evaluated various DL architectures, such as convolutional neural networks, vision transformers, generative adversarial networks, vision language models, and foundation models. Moreover, the study connects academic research with practical agricultural applications, providing guidance on the suitability of these models for production environments. This comprehensive review offers valuable insights into the current state and future directions of deep learning in plant disease detection, making it a significant resource for researchers, academicians, and practitioners in precision agriculture.

Polarization-controlled metasurface for simultaneous holographic display and three-dimensional depth perception

Abstract Simultaneous optical display and depth perception are crucial in many intelligent technologies but are usually realized by separate bulky systems unfriendly to integration. Metasurfaces, artificial two-dimensional optical surfaces with strong light–matter interaction capabilities at deep subwavelength scales, offer a promising approach for manufacturing highly integrated optical devices performing various complex functions. In this work, we report a polarization-multiplexed metasurface that can functionally switch between holographic display and Dammann gratings. By tailoring the incidence polarization, the metasurface can display high-quality holographic images in the Fresnel region or project a uniform spot cloud nearly covering the entire 180° × 180° transmissive space. For the latter, a projection and three-dimensional (3D) reconstruction experiment is conducted to elaborate the potential in retrieving 3D complex spatial information. The current results provide a prominent way to manufacture lightweight and highly-integrated comprehensive imaging systems especially vital for cutting-edge intelligent visual technologies.

A Comprehensive Review of Vision-Based Sensor Systems for Human Gait Analysis.

Analysis of the human gait represents a fundamental area of investigation within the broader domains of biomechanics, clinical research, and numerous other interdisciplinary fields. The progression of visual sensor technology and machine learning algorithms has enabled substantial developments in the creation of human gait analysis systems. This paper presents a comprehensive review of the advancements and recent findings in the field of vision-based human gait analysis systems over the past five years, with a special emphasis on the role of vision sensors, machine learning algorithms, and technological innovations. The relevant papers were subjected to analysis using the PRISMA method, and 72 articles that met the criteria for this research project were identified. A detailing of the most commonly used visual sensor systems, machine learning algorithms, human gait analysis parameters, optimal camera placement, and gait parameter extraction methods is presented in the analysis. The findings of this research indicate that non-invasive depth cameras are gaining increasing popularity within this field. Furthermore, depth learning algorithms, such as convolutional neural networks (CNNs) and long short-term memory (LSTM) networks, are being employed with increasing frequency. This review seeks to establish the foundations for future innovations that will facilitate the development of more effective, versatile, and user-friendly gait analysis tools, with the potential to significantly enhance human mobility, health, and overall quality of life. This work was supported by [GOBIERNO DE ESPANA/PID2023-150967OB-I00].

Computer Vision and Advanced Computational Algorithms for Risk Assessment and Performance Enhancement in Track and Field Teaching

This paper discusses the application of computer vision and advanced calculation algorithm in evaluating the teaching risk and teaching effect of track and field. Because of the inherent uncertainty and risk of PE and sports activities (especially track and field), it is necessary to establish an effective risk management mechanism. Using computer vision technology, this paper puts forward a method of analyzing and processing video data to detect and track moving objects, so as to identify potential risks in real time. This method not only improves the safety of students in track and field classes, but also provides valuable insights for improving teaching methods and reducing sports injuries. This paper discusses the background subtraction motion detection algorithm, which is very important for dynamic image modeling and shadow suppression, and can realize accurate motion state detection. The ultimate goal is to ensure the healthy development of school sports and optimize the teaching results of track and field sports.

Application Research of Vision-Guided Grinding Robot for Wheel Hub Castings

The combination of vision and robotic grinding technology provides robots with “visual perception” capabilities that enable them to accurately locate the area to be ground and perform the grinding tasks efficiently. Based on the rough grinding requirements for wheel hub burrs proposed by a casting company, this paper investigates the application of a vision-guided grinding robot in treating burrs on wheel hub castings. First, through vision system calibration, the conversion from pixel coordinate system to robot base coordinate system is implemented, thus ensuring that the subsequently extracted burr point coordinates can be correctly mapped to the robot’s operational coordinate system. Next, the images of the burrs on wheel hub castings are collected and processed. All the burr points are extracted by applying image algorithms. In order to improve grinding accuracy, a height error compensation model is established to adjust the coordinates of the 2D-pixel points; the coordinate error after compensation was reduced by 58.33%. Subsequently, the compensated burr point trajectories are optimized by utilizing an intelligent optimization algorithm to generate the shortest grinding path. Through experimental analysis of the relationship between spindle speed and surface roughness, a grinding trajectory simulation model is constructed, and the simulation results are integrated into the robot system. Finally, actual wheel hub burr grinding experiments are performed to validate the effectiveness and practicality of the proposed solution.

Research on Innovative Apple Grading Technology Driven by Intelligent Vision and Machine Learning.

In the domain of food science, apple grading holds significant research value and application potential. Currently, apple grading predominantly relies on manual methods, which present challenges such as low production efficiency and high subjectivity. This study marks the first integration of advanced computer vision, image processing, and machine learning technologies to design an innovative automated apple grading system. The system aims to reduce human interference and enhance grading efficiency and accuracy. A lightweight detection algorithm, FDNet-p, was developed to capture stem features, and a strategy for auxiliary positioning was designed for image acquisition. An improved DPC-AWKNN segmentation algorithm is proposed for segmenting the apple body. Image processing techniques are employed to extract apple features, such as color, shape, and diameter, culminating in the development of an intelligent apple grading model using the GBDT algorithm. Experimental results demonstrate that, in stem detection tasks, the lightweight FDNet-p model exhibits superior performance compared to various detection models, achieving an mAP@0.5 of 96.6%, with a GFLOPs of 3.4 and a model size of just 2.5 MB. In apple grading experiments, the GBDT grading model achieved the best comprehensive performance among classification models, with weighted Jacard Score, Precision, Recall, and F1 Score values of 0.9506, 0.9196, 0.9683, and 0.9513, respectively. The proposed stem detection and apple body classification models provide innovative solutions for detection and classification tasks in automated fruit grading, offering a comprehensive and replicable research framework for standardizing image processing and feature extraction for apples and similar spherical fruit bodies.